Scholarlink

<p>To promote a patient-centered, sustainable healthcare ecosystem, this paper explores how blockchain and Federated Learning (FL) might be integrated into the Medical Internet of Things (MIoT). A decentralized architecture for improving data security, privacy, and interoperability between healthcare systems is proposed in this article. The article handles the drawbacks of centralized Machine Learning (ML) techniques, which frequently jeopardize the privacy of sensitive medical data, by utilizing the collaborative character of FL and the security aspects of blockchain. ML models are jointly trained while protecting patient privacy by leveraging distributed MIoT data. To effectively anticipate diseases, the study uses a variety of approaches, including FL coupled with AdaBoost, Extra Tree Classifier, Decision Tree (DT), or Linear Discriminant Analysis (LDA). Extensive testing encompassing feature selection, data splitting, cross-validation, and hyperparameter tuning guarantees the effectiveness and confidentiality of the suggested methodology. To assess the efficacy of the technique, performance measures such as Accuracy, Balanced Accuracy (BA), Fowlkes-Mallows Index (FM), Matthews Correlation Coefficient (MCC), and Bookmaker Informedness (BI) are calculated. The findings demonstrate an improved level of accuracy in contrast with conventional techniques. The research represents a novel approach to medical data analysis by choosing the best algorithm using several evaluation metrics and then incorporating it into FL.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Empowering secure and sustainable healthcare through federated learning and blockchain synergies in a Medical Internet of Things

  • 期刊论文

  • 发表日期：

    2025年6月12日

查看全文
  • Jing Wang,
  • Sarbajit Manna,
  • Muammer Aksoy,
  • Arindam Sarkar,
  • Md Arafatur Rahman,
  • Abdulfattah Noorwali,
  • Kamal M. Othman,
  • Mohammed J. F. Alenazi

摘要

To promote a patient-centered, sustainable healthcare ecosystem, this paper explores how blockchain and Federated Learning (FL) might be integrated into the Medical Internet of Things (MIoT). A decentralized architecture for improving data security, privacy, and interoperability between healthcare systems is proposed in this article. The article handles the drawbacks of centralized Machine Learning (ML) techniques, which frequently jeopardize the privacy of sensitive medical data, by utilizing the collaborative character of FL and the security aspects of blockchain. ML models are jointly trained while protecting patient privacy by leveraging distributed MIoT data. To effectively anticipate diseases, the study uses a variety of approaches, including FL coupled with AdaBoost, Extra Tree Classifier, Decision Tree (DT), or Linear Discriminant Analysis (LDA). Extensive testing encompassing feature selection, data splitting, cross-validation, and hyperparameter tuning guarantees the effectiveness and confidentiality of the suggested methodology. To assess the efficacy of the technique, performance measures such as Accuracy, Balanced Accuracy (BA), Fowlkes-Mallows Index (FM), Matthews Correlation Coefficient (MCC), and Bookmaker Informedness (BI) are calculated. The findings demonstrate an improved level of accuracy in contrast with conventional techniques. The research represents a novel approach to medical data analysis by choosing the best algorithm using several evaluation metrics and then incorporating it into FL.